Characterization of a Dielectric Barrier Discharge in axisymmetric ...

89 International Journal of Plasma Environmental Science and Technology Vol.2, No.2, SEPTEMBER 2008
Characterization of a Dielectric Barrier Discharge in axisymmetric and
planar configurations -Electrical Properties-
Boni Dramane, Noureddine Zouzou, Eric Moreau and Gerard Touchard
Laboratoire d’Etudes Aérodynamiques (LEA) , Université de Poitiers, ENSMA, CNRS
2 Bd. Marie & Pierre Curie, Téléport 2, BP. 30179, 86962, Futuroscope Chasseneuil Cedex, France
Abstract— The characteristics of a Dielectric Barrier Discharge (DBD) in axisymmetric and planar configurations are
studied and compared by measuring their electrical discharge parameters and the airflow influence on these parameters.
With the axisymmetric reactor, the application of AC high voltage generates a stable and quasi-diffuse discharge whereas
the planar reactor generates a filamentary and unstable discharge. When the flow rate increases the discharge current and
the electric power decrease in both cases. This effect is less pronounced in the planar configuration.
Keywords— Dielectric barrier discharge, Wire to cylinder reactor, Plane to plane reactor
I. INTRODUCTION
Dielectric Barrier Discharge (DBD) is well known
to produce highly non-equilibrium plasmas in a
controllable way at atmospheric pressure. This method
can effectively generate atoms, radicals, and excited
species with energetic electrons at moderate gas
temperatures [1, 2].
Because of its high electron density and energy, the
DBD in air has been used to clean hazardous air
pollutants (HAPs) [3-5].
In general, operating at atmospheric pressure, the
DBD is generated in filamentary form [6]. DBDs can
also be obtained in a diffuse form under certain
configurations and conditions [7-9].
In order to remove ultra fine particles from Diesel
exhaust, two DBD reactors have been used to determine
which of them gives the lowest energy density for an
efficient removal. The reactors are in axisymmetric and
planar configurations.
This preliminary study was performed by measuring
the electrical parameters of the discharge: applied voltage,
discharge current, transferred charge, and electric power
consumption. The results obtained in both configurations
were compared. A detailed attention was given to the
influence of the flow on the discharge behavior.
Figure 1 shows a schematic illustration of the
axisymmetric reactor used in this work, while figure 2
shows the planar one.
Copper sheet
Plasma
1 2
22 nF 100 Ω
φ 20 mm
Air IN
φ 0.2 mm
2 mm
Pyrex tube
80 mm
Fig. 1. Schematic of the axisymmetric reactor
Pyrex plates
Grounded
electrode
High Voltage
electrode
High Voltage
electrode
II. EXPERIMENTAL
Air IN
2 mm
2 mm
In order to determine which type of configuration
will make it possible to obtain a diffuse discharge or
rather a filament discharge, we used two types of reactor
in different configurations.
Corresponding author: Boni Dramane
e-mail address: boni.dramane@lea.univ-poitiers.fr
Aluminium
sheets
Plasma
Grounded
1 2
electrode
22 nF 100 Ω
2 mm
Received: July 10th, 2007, Revised: March 12th, 2008, Accepted:
March 30th, 2008
Fig. 2. Schematic of the planar reactor.

Dramane 90
The axisymmetric configuration setup consists of a
Pyrex tube, a copper sheet, a stainless wire and an air
circulation system. Outer diameter of the Pyrex tube was
20 mm and the thickness was 2 mm. The copper sheet
with 8 cm width was rolled up on the outer surface of the
Pyrex tube. The stainless wire of 0.2 mm of diameter was
maintained tended on the central axis of the tube and
connected to the high voltage.
The planar configuration setup consists of two
aluminum sheets, two plates made of Pyrex and an air
circulation system. With a thickness of 0.1 mm and an
area of 50 x 47 mm 2 , each aluminum foil is stuck on the
dielectric plate. Using spacers made of glass of 2 mm
thick, the active volume of the discharge was 50x40x2
mm 3 . The top sheet, used as the active electrode, was
connected to high voltage and the bottom one was
grounded. To avoid the edge effect, the electrode
contours were encapsulated in epoxy resin.
The high voltage power amplifier (Trek, 20/20C,
±20 kV/ ±20 mA) was driven by a function generator
(HAMEG, HM8130) to generate high voltage sine wave.
High voltage up to 20 kV was applied to the high voltage
electrode with frequencies of 0.5, 1 and 2 kHz. Applied
voltage was monitored with an oscilloscope (Tektronix,
TDS3014B) with a high voltage divider (Lecroy 1000:1,
20 kV). The total current of discharge is measured with a
resistor of 100 Ω and the transferred charge with a
capacitor of 22 nF.
All the experiments were carried out under
atmospheric pressure at room temperature. The airflow
inside the reactors was controlled. The flow rate (Q) is
measured with a flow meter (Brooks, tube size R-6-15-A,
Tantalum ball).
pulses dominate the negative half-cycle. This behavior is
also observed in point to plane DBD [10].
The transferred charge waveform shown in figure 3-
b presents an evolution which can be break up into two
steps. During the positive half-cycle, the transferred
charge increases and after the voltage maximum it
Current (mA)
Charge (nC)
20
10
0
-10
-20
-20
0,0 0,5 1,0 1,5 2,0
Time (ms)
500
250
0
-250
(a)
20
10
0
Voltage (kV)
-10
20
10
0
-10
Voltage (kV)
III. RESULTS AND DISCUSSION
A. Discharge characteristics in each configuration
Typical oscillograms of the applied voltage,
discharge current, transferred charge and charge-voltage
curve in the Lissajous figure with axisymmetric and
planar reactors are shown in figures 3 and 4, respectively.
Figure 3- a shows the time evolution of the applied
voltage and the discharge current obtained from the
measured current. Indeed, the measured current includes
fast component (current pulse) and slow component
(capacitive and pseudo continuous current). The
capacitive current, which was estimated from the I-t
curve with applied voltage low enough to generate any
discharge, was subtracted from the measured current to
obtain the discharge current. All the discharge currents
hereafter are presented in this way.
The discharge current waveform shows only few
pulses of current on the positive half-cycle, while on the
negative one the number of pulses is significant. In the
positive half-cycle of the applied voltage, the plasma is
characterized by a glow-like type. However, the Trichel
Charge (nC)
-500
-20
0,0 0,5 1,0 1,5 2,0
500
250
0
-250
Time (ms)
-500
-20 -10 0 10 20
Voltage (kV)
(b)
(c)
Fig. 3. Axisymmetric reactor characteristics: (a) time evolution of the
discharge current, (b) time evolution of the transferred charge, and (c)
charge-voltage curve. Conditions: V peak = 12 kV; f = 1 kHz; Q = 4.7
L.min -1 .

91 International Journal of Plasma Environmental Science and Technology Vol.2, No.2, SEPTEMBER 2008
stabilizes. The transferred charge decreases again for
another sequence.
The charge-voltage curve (figure 3- c) represents a
close loop, which will be described in the following
section.
In contrast, the discharge in planar configuration is
characterized by a filamentary behavior (figure 4- a). The
current pulses are higher in magnitude and less numerous
during the positive half-cycle than on the negative one.
The time evolution of the transferred charge shows
similar steps as that found in axisymmetric configuration
(figure 4- b).
Another difference between both configurations is
their charge-voltage curve. It is noted that, in opposition
to the planar configuration, with a figure close to a
parallelogram [3], the charge-voltage curve in
axisymmetric configuration shows a kind of ring of the
shape of an ellipse flattened on the edges.
The slope of the curve represents the total
capacitance (C tot ) of the reactor [11]. Indeed, the total
capacitance of the reactors without discharge is
1/C tot =1/C d +1/C g , where C d is the dielectric capacitance
and C g is the air gap capacitance.
In the planar configuration, the filaments make a
short-circuit in the gap when the discharge occurs. Then,
the total capacitance, which is equal to the capacitance of
the dielectric, becomes larger.
However, in the axisymmetric configuration, the
discharge occurs around the wire without making a shortcircuit
in the air gap. The expansion of the discharge in
the gap varies with the voltage, inducing a variable gap
capacitance and obviously a variable total capacitance.
The slope of charge-voltage curve increases more slowly
than in the planar configuration, which explains the
flattening of the edges.
of the frequency from 1 kHz to 2 kHz for a fixed voltage
of 12 kV involves also an increase of the number of
discharge current pulses.
In contrast, the charge-voltage curve remains nearly
unchanged on figure 7- b compared to figure 4- c where
V peak = 12 kV and f = 1 kHz.
Current (mA)
300
150
0
-150
-300
-20
0,0 0,5 1,0 1,5 2,0
Charge (nC)
500
250
0
-250
Time (ms)
(a)
20
10
0
-10
20
10
0
-10
Voltage (kV)
Voltage (kV)
B. Effect of applied voltage and frequency on discharge
characteristics
Figures 5- a and 5- b show the time evolution of
applied voltage, discharge current and charge-voltage
curve for axisymmetric reactor when the applied voltage
is V peak = 16 kV and the frequency is f = 1 kHz.
Comparatively to figure 3 where the conditions
were V peak = 12 kV and f = 1 kHz, discharge current
increases as well as the number of Trichel pulses. In
addition, a glow-like discharge component appears on the
negative half-cycle. The charge-voltage curve increases
with the applied voltage.
Increasing the frequency at 2 kHz (figure 6- a)
implies the increase of the magnitude and the number of
discharge current pulses. A glow-like discharge
component can also be observed during the negative half
cycle. The charge-voltage curve decreases with the
frequency (figure 6-b).
In the planar configuration, when the voltage is
raised to V peak = 16 kV at fixed frequency of 1 kHz, the
number of discharge current pulses increases. In addition,
the charge-voltage curve increases (figure 7- a). The rise
Charge (nC)
-500
-20
0,0 0,5 1,0 1,5 2,0
Time (ms)
(b)
500
250
0
-250
-500
-20 -10 0 10 20
Voltage (kV)
(c)
Fig. 4. Planar reactor characteristics: (a) time evolution of the
discharge current, (b) time evolution of the transferred charge, and (c)
charge-voltage curve. Conditions: V peak = 12 kV; f = 1 kHz; Q = 4.7
L.min -1 .

Dramane 92
20
20
500
10
10
250
Current (mA)
0
-10
0
Voltage (kV)
Charge (nC)
0
-10
-250
-20
-20
0,0 0,5 1,0 1,5 2,0
Time (ms)
(a)
-500
-20 -10 0 10 20
Voltage (kV)
(b)
Fig. 5. (a) Time evolution of the discharge current. (b) Charge-voltage curve for axisymmetric reactor.
Conditions: V peak = 16 kV; f = 1 kHz; Q = 4.7 L.min -1 .
20
20
500
10
10
250
Current (mA)
0
0
Voltage (kV)
Charge (nC)
0
-10
-10
-250
-20
-20
0,00 0,25 0,50 0,75 1,00
Time (ms)
(a)
-500
-20 -10 0 10 20
Voltage (kV)
(b)
Fig. 6. (a) Time evolution of the discharge current. (b) Charge-voltage curve for axisymmetric reactor.
Conditions: V peak = 12 kV; f = 2 kHz; Q = 4.7 L.min -1 .
500
500
250
250
Charge (nC)
0
Charge (nC)
0
-250
-250
-500
-20 -10 0 10 20
Voltage (kV)
(a)
-500
-20 -10 0 10 20
Voltage (kV)
(b)
Fig. 7. Charge-voltage curve for planar reactor.
(a) Conditions: V peak = 16 kV; f = 1 kHz; Q = 4.7 L.min -1 . (b) Conditions: V peak = 12 kV; f = 2kHz; Q = 4.7 L.min -1 .

93 International Journal of Plasma Environmental Science and Technology Vol.2, No.2, SEPTEMBER 2008
C. Effect of air flow on discharge characteristics
The air circulation system consists of a pressure
reducing valve, a decicator containing CaSO 4 and a flow
meter. The air used is provided by the compressed air
network.
Several flow rates were applied to the discharge but
only the two extreme values will be shown, i.e. 1.6
L.min -1 and 15.8 L.min -1 . The flow is introduced inside
the reactor first, then the voltage is quickly raised up to
the desired value and the discharge is maintained during
one minute and finally the discharge is stopped.
Figures 8 and 9 show the waveforms of the chargevoltage
curve obtained with two flow rates in the
axisymmetric and planar configurations, respectively.
When the flow rate is increased, the discharge current
decreases and the ignition voltage of the discharge
increases (figure 10). Consequently, with higher flow
rates the transferred charge is lower. This effect is more
pronounced in the axisymmetric configuration.
The effect of the flow rate on the average power is
shown on figure 11. The flow rate varies between 1.6 and
Charge (nC)
500
250
0
-250
1,6 L.min -1
15,8 L.min -1
Courant(mA)
Fig. 10. Axisymmetric reactor characteristics: time evolution of the
discharge current. Conditions: V peak = 12 kV; f = 1 kHz.
Average power (W)
20
10
0
-10
-20
-20
0,0 0,5 1,0 1,5 2,0
60
40
20
0
1,6 L.min -1
15,8 L.min -1
Time(ms)
Axisymmetric reactor ( 1,6 L.min -1 )
Axisymmetric reactor (15,8 L.min -1 )
Planar reactor ( 1,6 L.min -1 )
Planar reactor (15,8 L.min -1 )
8 10 12 14 16 18 20
Voltage (kV)
Fig. 11. Average power/voltage curve for axisymmetric and planar
reactors at f = 1 kHz.
20
10
0
-10
Voltage (kV)
-500
-20 -10 0 10 20
Voltage (kV)
Charge (nC)
Fig. 8. Charge-voltage curve for axisymmetric reactor. Conditions:
V peak = 12 kV; f = 1 kHz.
500
250
0
-250
1,6 L.min -1
15,8 L.min -1
-500
-20 -10 0 10 20
Voltage (kV)
Fig. 9. Charge-voltage curve for planar reactor, Conditions: V peak = 12
kV; f = 1 kHz.
15.8 L.min -1 . A high flow rate induces a low average
power for both reactors. However, this tendency is more
important with the axisymmetric reactor than the planar
one.
During the discharge, it seems that some charges
are transported outside the plasma section because of the
airflow. This phenomenon is pronounced at high flow
rates and explains the result shown in figures 8-11. In
addition, the streamers at the origin of the filamentary
discharge progress rapidly in the gap (about 10 ns). Thus,
the filamentary discharge is not very affected by the
velocity of the air (some m/s).
IV. CONCLUSION
The characteristics of the Dielectric Barrier
Discharge in axisymmetric and planar configurations are
studied and compared by measuring their electrical
discharge parameters and the airflow influence on these
parameters.

Dramane 94
It was observed that the DBD discharge
characteristics in axisymmetric configuration are
different from those of the planar configuration. By
modifying the geometry of the reactors, a different
discharge type can be obtained using the same electric
excitation.
In axisymmetric configuration, DBD discharge is
rather quasi-diffuse while, in the planar one, the
discharge is filamentary.
The study of the influence of the air flow on the
discharge show that discharge current, transferred charge
and average power decrease with the flow rate. This
evolution is more pronounced with the axisymmetric
reactor.
Now, both reactors are going to be tested for
particle precipitation, agglomeration or destruction.
REFERENCES
[1] F. S. Denes and S. Manolache, “Macromolecular plasmachemistry:
an emerging field of polymer science”, Prog. Polym.
Sci., Vol. 29, No.8, pp. 815-885, August 2004.
[2] N. Gherardi and F. Massines, “Mechanisms controlling the
transition from glow silent discharge to streamers discharge in
nitrogen”, IEEE Trans. Plasma Sci., Vol. 29, No.3, pp. 536-544,
2001.
[3] J. H. Byeon, J. Hwang, J. H. Park, K. Y. Yoon, B. J. Ko, S. H.
Kang and J. H. Ji, “Collection of submicron particles by an
electrostatic precipitator using a dielectric barrier discharge”, J.
Aerosol Sci., Vol. 37, pp. 1618-1628, November 2006.
[4] S. Sato, M. Kimura, T. Aki, I. Koyamoto, K. Takashima, S. Katsua
and A. Mizuno, “Removal of diesel particles using an electrostatic
precipitator with a barrier discharge electrode as a dust pocket”,
Proc. ESA/IEJ/IEEE-IAS/SFE joint Conference on Electrostatics
2006, Berkley California, pp. 754-762, 6-9 June, 2006.
[5] Y. Dan, G. Dengshan, Y. Gang, S. Xianglin and G. Fan, “An
investigation of the treatment of particulate matter from gasoline
engine exhaust using non-thermal plasma”, J. Hazard. Mater., Vol.
B127, pp. 149-155, December 2005.
[6] Z. Fang, Y. Qiu, C. Zhang, and Kuffel, “Factors influencing the
existence of the homogeneous dielectric barrier discharge in air at
atmospheric pressure”, J. Phys. D: Appl. Phys., Vol. 40, pp. 1401-
1407, February 2007.
[7] J. R. Roth, J. Rahel, X. Dai and D. M. Sherman, “The physics and
phenomenology of One Atmosphere Uniform Glow Discharge
Plasma (OAUGDP TM ) reactor for surface treatment applications”, J.
Phys. D: Appl. Phys., Vol. 38, pp. 555-567, February 2005.
[8] J. Rahel and D. M. Sherman, “The transition from a filamentary
dielectric barrier discharge to a diffuse barrier discharge in air at
atmospheric pressure”, J. Phys. D: Appl. Phys., vol. 38, pp. 547-
554, February 2005.
[9] M. Abdel-Salam, A. Hashem, A. Yehia, A. Mizuno, A. Turky and
A. Gabr,“Characteristics of corona and silent discharges as
influenced by geometry of the discharge reactor”, J. Phys. D: Appl.
Phys., Vol. 36, pp. 252-260, January 2003.
[10] M. Petit, A. Goldman and M. Goldman, “Glow currents in a pointto-plane
dielectric barrier discharge in the context of the chemical
reactivity control”, J. Phys. D: Appl. Phys., Vol. 35, pp. 2969-2977,
November 2002.
[11] I. Nagao, M. Nishida, K. Yukimura, S. Kambara and T. Maruyama,
“NO x removal using nitrogen gas activated by dielectric barrier
discharge at atmospheric pressure”, Vacuum, Vol. 65, No. 3-4, pp.
481-487, May 2002.